Abstract: A hybrid crude oil scheduling optimization model integrating mathematical programming with heuristic algorithms is proposed to tackle the complexity arising from intricate operational constraints, dynamic planning requirements, and limited tankage in refining enterprises. The model employs a two-stage serial framework: a mathematical programming-based crude oil processing planning model determines optimal feed schedules for two distillation units under both general and enterprise-specific constraints; subsequently, a heuristic daily scheduling model generates detailed operational schemes based on the optimized feed plan. Validation results show that the optimized feed schedule satisfies all unit and route-specific constraints. The derived scheduling scheme adheres to tank compatibility and pipeline transportation restrictions, with 84.75% and 88.37% of tank operations in the incoming and feed tank farms lasting over 6 hours, respectively, and 83.33% of main crude transfers maintaining a minimum 5hours settling time. All performance metrics meet customized operational requirements, demonstrating the practical viability of the proposed approach.

Key words: crude-oil scheduling, intelligent scheduling, scheduling optimization, mathematical programming, heuristic algorithm